Some of his testing methods are unreliable for what he is attempting to prove.
As an example, his radiated susceptibility test (Unshielded vs. Shielded vs. Shielded and ungrounded).
He is using what, a network hotspot as a means to determine if radiated emissions will get into his cable? What is the max transmission power on that thing, 0.5W at 2 GHz? From that he ascertains that there is no effect on grounding the shield.
I can tell you for a fact that radiated susceptibility is a real thing. A long unshielded cable run will pick up emissions and electrical noise. Whether it is 60Hz hum from fluorescent lighting or emissions from other equipment. It may not happen in the equivalent of 5 feet of cable, but if you look at 50 or 100' of cable passing through an area with a high electromagnetic field or high RF emissions, you will see those on the signal. Also his testing by using an oscillioscope set to time display is retarded. He should be using a spectrum analyzer to look for the peaks from his radiated source, or a high speed O-scope displaying an eye diagram, not that a spread spectrum WiFi hotspot is likely to produce much of anything on either of those instruments.
Also his conclusion on ground loops is incorrect. I have seen systems with ground potential variations of >5V. Grounding the cable shield on both ends generated a fair amount of current flow and played havoc with the signals passing through those cables. Ground on one end of the cable only.
Modern differential signals such as Ethernet are designed to be tolerant of things like radiated RF or electromagnetic emissions.
https://en.wikipedia.org/wiki/Differential_signaling
So his testing specifically against this type of signal is going to lead to his conclusions. While radiated noise may not be such a huge problem for modern Ethernet, a general statement that it is not a problem, at all, is blatantly incorrect. Even the differential signals on Ethernet will eventually get overwhelmed when the EMF gets large enough. I worked at a facility with a 3000HP electric motor on a variable frequency drive. When that motor was running, non-shielded Ethernet passing through that area would not work, at all. The shielded cable (only on one end) worked fine and no difference in transmission speed was seen with the motor running or not.